At OkayAfrica, we’re constantly discussing the impact of AI. Whether it’s how we use it in our own work or how we should cover its intersections with the subjects we report on across Africa, it’s a conversation we keep returning to. Music, in particular, brings those debates to the forefront. AI raises big questions about creativity, ownership, labor, and access, and there are still no clear answers.
So I decided to speak to someone who is knee-deep in the world of AI and African music, and who could speak to some of the tensions, possibilities, and ethical questions shaping it in real time.
For Lagos-based software developer and former music producer Philip Olajide-Philips, those questions are both theoretical and personal. When he couldn’t afford the studio time he needed, he built an alternative. His startup, Korin AI (korin means “to sing” in Yoruba), is an AI music platform that allows users to generate songs in African languages and accents. It is, in some ways, a direct competitor to larger platforms like Suno AI and Udio that are currently upending the music industry. But as Olajide-Philips tells OkayAfrica, Korin AI’s value proposition is that it is being built with African languages, accents, and creative realities in mind.
“Most of the AI music tools out there were built for the Western world,” he says. “If you put Yoruba or Zulu lyrics into them, they’ll sing the words, but they pronounce them like British or American English. With Korin, the goal is that it sounds like an African voice in your own language, not like a foreign singer reading your words.”
With a new 2.0 version debuting in May, Korin AI is geared toward emerging and independent artists, producers, filmmakers, and anyone with musical ideas but limited access to professional studios. In many African cities, recording a polished single can cost hundreds of dollars and take weeks to complete. Olajide-Philips sees the platform as a way to cut down both time and cost, while still keeping human creativity at the center of the process.
“I saw that about 75% of upcoming artists in Africa simply can’t afford top studios,” Olajide-Philips says. “You can be very talented, but if your production quality is low, people won’t listen. So I asked myself: in this AI world, why can’t we have a virtual African studio that anyone can access?”
His question reflects the larger debate around AI music itself. At a time when many AI platforms are facing criticism for scraping artists’ work without consent, Olajide-Philips says Korin is trying to build differently. Rather than pulling audio from the open internet, the company relies on partnerships with studios and artists, paid recordings, and royalty or revenue-share agreements. The bet is that an Africa-first, more ethical data model will be the right approach, with a competitive advantage.
So far, Korin AI has drawn around 1,500 users on the first version without significant marketing. Backed largely by Olajide-Philips, his co-founder Solomon Ayodele Ogunbowale, and early support from First Founders, the team is holding off on a bigger push until Korin 2.0 is ready.
In the conversation below, Olajide-Philips speaks to OkayAfrica about why he doesn’t believe AI will replace human creativity, what it means to build African-native models in a landscape where most tools still sound distinctly Western, and where the future of Korin of music in this new age of AI.
This interview has been edited for length and clarity.
OkayAfrica: When I talk to people about AI, I see two camps: one says, “AI is the future, get on board,” and the other says, “AI is an existential threat, the world is ending.” Where do you land on that spectrum, especially when it comes to music?
Olajide-Philips: AI has come to stay and there is nothing that is going to change that. Anytime a new technology comes, people always say no at first. It happened with GSM when it first came to Africa, it happened with the internet, it happened with WhatsApp. People resist because it threatens the way they’ve been doing things, or they’re afraid they’ll lose their jobs. But you can’t stop the advancement of technology. After a while, everybody gets used to it. I see AI the same way.
How did your own frustration with music production turn into building Korin AI?
In 2024, I wanted to produce something and the studios I contacted were too expensive and too slow. I’m a software developer, I understand AI, and I have experience in music, so I thought: why can’t I do this faster and cheaper?
I tried the AI music tools that existed and realized most of them were built for the Western world. When I put my song into them, the AI was trying to sing like a white man, not in my dialect. It was singing, but not like us. I couldn’t find any specific model designed for African data. I even asked myself, ‘What’s wrong with these guys? Do they want to remove Africa from the whole mainstream?’
So I ran a proof of concept. I told some founders I knew, ‘Send me your company details, I’ll create jingles for you.’ I ended up doing about 20 jingles in two days. Some people thought I’d gone into a professional studio. When they heard the results, they were blown away. That’s when I knew this could be a real product.
You’ve positioned Korin AI as a kind of music lab for African creators. What does that mean in practice, and what makes it different from the global tools already out there?
Most of the AI tools out there were built with the West in mind. If you input lyrics in Yoruba, Zulu, or Igbo, they will sing the words, but they pronounce them in the language the model was trained on like British or American English. So you hear it, and it’s not sounding like Yoruba or Zulu, it’s sounding like someone reading your language with a foreign tongue.
Korin AI is trained with African datasets, built by Africans, and designed with Africans in mind. When you put your lyrics in, the goal is that it sounds like an African voice in your own language, not like a British or American singer reading unfamiliar words.
Right now, our deepest data is in Nigeria and a few other languages we’ve sampled. We’re starting there and plan to expand to Ghana, Kenya, Zimbabwe, Lesotho, Gambia, and beyond, partnering with local producers and artists, collecting data, and updating our models so that when someone from those countries uses Korin, it actually sounds like home.
A lot of AI platforms are criticized for scraping the internet for training data. At the same time, one worry I have is that AI could flatten Africa into a single sound.
We’ve been intentional about that from day one. We don’t want to exploit anyone. Those guys abroad, they just scrape the internet for data and don’t pay people. We don’t want to do that.
Right now, we work through partnerships with existing production houses in Nigeria. Some we pay directly. With others, we have revenue-share agreements, so as we make money, they earn a percentage.
We also bring in volunteers to studios and record their voices. Some are paid, others sign lifetime licenses. But in every case, it’s consented and clearly agreed. Our long-term vision is that you could come to Korin AI and say, ‘I want to create a song with a specific artist’s voice,’ and because we already have a license and royalty arrangement with that artist, it’s legal and they get paid. It’s not about scraping and hoping no one notices.
Korin AI is just starting, so we don’t pretend to cover everything yet. Right now our strongest base is Nigeria, plus a few other African languages we’ve already sampled. The goal is to expand to all major African languages over time by partnering with local producers and individuals, collecting data, and training the models with those voices, accents, and musical styles.
The product is already live, and people in places like Lesotho, Gambia, or Kenya can use it, but it may not be as deep or as accurate yet as it is for Nigerian languages. Funding really determines how quickly we can go deep into each language and region.
Many are worried that tools like Korin will reduce demand for human creative work. What do you say to them?
AI is not taking anybody’s job, but the job of those that are afraid of AI will be taken by those that are trained and using AI tools.
Korin is not here to replace producers. It’s here to augment them. Work that used to take weeks can now be done in a day or two. Producers can generate multiple versions quickly, offer clients more options, and deliver faster while still adding value.
So AI removes the drudgery, but it doesn’t remove human oversight, taste, and creativity. Those are still essential.
We recently reported on a 21-year-old Rwandan artist, Elvin Cena, who used Suno AI to rework a song he didn’t like. It blew up. He felt conflicted because this 30-minute AI version became his biggest song. Is that still his work? And do you think artists should disclose when AI has been used?
In my opinion, yes, it’s 100% his work, especially if he was on a paid subscription and has the rights.
Human creativity is still at the core. He wrote the lyrics, chose the genre, the melody, and tweaked the prompts. The AI can’t do that by itself. It needs a human in the loop. It’s the same with my code. Sometimes AI helps me write it in 15 minutes where I used to spend weeks, but I still ask: who wrote the code? The creativity is still from the human being.
Some people will worry that if you can generate a song quickly, it will devalue music. But you can also make very bad songs with AI. A good song is still a good song. The value comes from human creativity.
I like the idea of a co-created credit style. The way we already say ‘composed by X, arranged by Y, mixed and mastered by Z,’ we can add AI into that chain, almost like another producer. But for me, the most important thing is ownership. If you wrote the lyrics, chose the genre and melody, and directed how the AI should behave, then you should claim the work. The AI is a tool… it is not the artist.
Platforms like Suno AI are already popular globally. If big Western players decide to go hard on African languages too, how do you compete?
There’s always competition in any business. The big Western companies look for markets where they can easily get their money back. Historically, Africa hasn’t been their first priority. That’s why their models launched without African data in the first place. Our strength is that we’re indigenous. We know the languages, the accents, the people, even in rural areas. They can come later and try to catch up, but they still won’t be as deep and native as someone who’s built from inside the culture. When people need an authentic African sound in AI, I want them to say, ‘Go to Korin AI.’
Looking ahead five years, what would success look like for Korin AI?
In five years, we want at least 10 million users on the platform. And when people anywhere in the world say ‘African AI music,’ I want Korin AI to be the first name they think of; the place you go to generate any native African song, in a truly native voice.
